The release of Apple’s new MacBook Pro with Retina Display, and the discussion regarding its proprietary, difficult to expand nature (as found by iFixit when they tore down the MBPR), has triggered some thoughts on where personal computing is headed.
For personally-owned computers, many manufacturers are chasing Apple in one of two ways – either making media tablets, to try to piggyback on the iPad’s success, or making “ultrabooks”, to try to piggyback on the MacBook Air’s success. Sometimes, they’re making weird combinations of the two. In addition, Windows 8 seems to be aimed more at media tablets than at the conventional desktop and laptop computers, as far as user interface goes.
I think that this trend will continue, but with media tablets and ultrabooks merging into two form factors of the same device (including some devices that can convert between the two, or can have a keyboard added, ala the Asus Transformer family) – low to moderate processing power, very light weight, very thin, no or few moving parts (fans at most), and disposable. I’m guessing that the name will change, or it’ll just be “tablets” and “notebooks”, the differentiating factor being whether it has a keyboard integrated. Some will be expandable (primarily by adding SD cards or equivalent), some won’t. I think touch will be present on the personal notebooks of the future, although the keyboard will definitely survive, and another pointing device will be there.
For operating systems, most likely, the one full-screen task at a time usage model of a media tablet will stick – most people run all of their programs full screen anyway. Curated software repositories have been a wild success with users, and I don’t see that changing. Some platforms will be more open about installing software that isn’t in their repositories than others, but installing third party OSes may be a no-go (depends on whether anti-trust suits are thrown around, really). Also, cloud services will continue their popularity, relying on servers for the heavy lifting. I don’t think that the desktop model as it is today will survive on personally-owned computers, and I’m not convinced that that’s a bad thing for most users.
As far as gaming, some may be done on these machines, but there will either be dedicated consoles, or services like OnLive using server farms.
In business, tablets and “bring your own device” seem to be growing trends, meaning that for many users, the personally-owned computers will also be what they do work on. Businesses will almost certainly make heavy use of cloud services, albeit run in their own datacenters in many cases.
However, for some heavy lifting, the conventional desktop workstation will still stick around. Blade workstations may be popular, where many workstations are in a datacenter, and then a lighter device connects as a dumb terminal to that machine. However, the limited I/O of a blade workstation tells me that true desktop workstations will be available for the applications that need the power. In either case, these workstations will essentially be server hardware with a GPU, and due to potentially very different economies of scale from today, this stuff could be rather expensive – as in, the old days of when a workstation was dedicated RISC hardware. Mobile workstations will also still exist for the applications that won’t work with the blade approach, and need to be mobile, but they’ll be similar to how they were in the early 1990s – as expensive as a good car due to their very poor economies of scale, likely made from a mix of blade server components (poorly suited to the job) and ultrabook components, and extremely bulky.
As far as what computing enthusiasts use, there’ll be a mix of choices. Many enthusiasts will use the media tablets or ultrabook-style devices, potentially hacked to run third-party OSes. However, given the emphasis on cloud computing that I’m predicting, that may not be enough. I predict that some enthusiasts will run servers of their own, and then point their tablets or ultrabooks at those servers. In addition, for enthusiasts that are nostalgic for the old ways of running a desktop, and get workstations. The extremely rich enthusiasts will use mobile workstations.
Now, for us enthusiasts, this sounds like a dystopia – the power being taken from the user, put in the control of the megacorporations that make our computers. That said, it doesn’t necessarily have to suck. Like I said, running our own cloud services at least keeps control of our data, and there’ll be plenty of surplus servers to make it possible. The iOS jailbreaking community and the Android third-party ROM community show that we’ll be able to own our own devices, too, and for many workloads, they will be powerful.
There, however, are two unanswered questions.
First, how will people who aren’t already experienced with computers begin to explore the system?
Second, how will development occur? After all, right now, development for every mobile platform occurs on a desktop platform.
For the second problem, I predict that the OS makers will release IDEs that run on the mobile platform – they can’t afford to alienate developers by requiring them to get an expensive workstation, because they’re dependent on having a healthy software base. (While some will cite Apple as a modern-day example, please realize that I’m referring to very different price classes here. For comparison, I predict a loaded Mac Pro to be on the low end of the price scale for workstations.) However, this will require that the system not be locked down.
The answer may be either “Developer Edition” devices that are already unlocked and can run unsigned software, solving the second problem, or a “developer switch” that’s requires a “special dance” of sorts to get to – not explicitly disallowed, just intentionally hard to get to. The switch approach, used by webOS (webos20090606) and Maemo (Red Pill mode), is my favorite approach for keeping normal users out, but allowing someone who’s willing to accept the consequences access, without requiring them to defeat the system security.
Of course, this is assuming that we’re still in a “constant growth” economy in 10 years. If we aren’t (and we shouldn’t be, but there are reasons that the powers that be want constant growth economies), throw all these predictions out the window, computers will be very expensive in general, and far more similar to workstations as far as longevity and expandability, although there may be heavy standardization of the platform to improve economies of scale. In that case, I do think OSes in a similar style to iOS and Android will be successful and widely popular, but lock-down of what OS you can load won’t be tolerated due to the need to make computers non-disposable.